Risk of severe life threatening asthma and beta agonist type: an example of confounding by severity.

BACKGROUND A study was undertaken to test the hypothesis that a particular inhaled beta agonist, fenoterol, increases the incidence of severe life threatening asthma. METHODS A retrospective cohort was assembled comprising 655 patients with asthma aged 15-55 years who attended a single Auckland hospital for acute asthma between 1 January 1986 and 31 December 1987 (the "index event"). Patients were followed for the occurrence of death from asthma or admission to the intensive care unit for asthma, until death or 31 May 1989. Data on asthma medications and asthma severity were obtained from forms used specifically for managing patients with acute asthma in the emergency department and maintained as part of the hospital record and/or from the hospital record (when patients were admitted). RESULTS Following the index event 90 admissions to the intensive care unit (ICU) and 15 asthma deaths were identified. Before adjusting for asthma severity, patients using inhaled fenoterol had a greater incidence of severe life threatening asthma than patients using inhaled salbutamol (RR = 2.1, 95% CI 1.4 to 3.1). After controlling for two markers of severe asthma used in previous studies-a hospital admission in the previous year and prescribed oral corticosteroids-the relative risk estimate declined to 1.5 (95% CI 1.0 to 2.3). After controlling further for the number of hospital admissions during the study period, continuous oral corticosteroid use rather than short courses of treatment, severity of the previous attack requiring a hospital visit, and race, fenoterol was not associated with severe life threatening asthma at the time of attendance for a previous hospital visit (RR = 1.0, 95% CI 0.6 to 1.7). CONCLUSION Fenoterol is used more often by patients with severe asthma and, after adjusting for differences in baseline risk, it does not increase the risk of severe life threatening asthma.